Generation of real-time SDTM datasets and metadata through a generic SDTM converter mechanism CDISC (CDASH/SDTM) integration into OC/RDC Peter Van Reusel Business Unit Director CRO Services Business & Decision Life Sciences Tel Fax Mobile Sint-Lambertusstraat 141 Rue Saint- Lambert 1200 Brussels

Agenda 1 Introduction 2 CDASH/SDTM in OC/RDC 3 SDTM Converter Mechanism 4 Conclusion

Agenda 1 Introduction 2 CDASH/SDTM in OC/RDC 3 SDTM Converter Mechanism 4 Conclusion

CDISC models AcronymTitleUse Clinical Data Acquisition Standards Harmonization Data Collection Study Data Tabulation Model Content Analysis Dataset ModelContent Case Report Tabulation Data Definition Specification Metadata ODM Operational Data ModelFormat LAB Clinical Laboratory ModelContent PRG Protocol Representation Group Content BRIDG/HL7BRIDG/HL7 SDTM ADaM Define.xml CRT-DDS CDASH

Near Term Approach: Legacy Data Conversion

Longer Term Approach: Global CDISC Integration

Agenda 1 Introduction 2 CDASH/SDTM in OC/RDC 3 SDTM Converter Mechanism 4 Conclusion

CDISC integration objectives To integrate global CDISC standards in OC/RDC – Use CDISC nomenclature across the CDM system Ensure metadata consistency Use of CDISC controlled terminology Avoid additional data conversion steps – Generate SDTM data as of study start Production of real-time SDTM deliverables Facilitate generation of ADaM datasets based on SDTM Allow CDISC consistency checking early in the process

CDISC integration prototype scope Current scope: Special purpose domains: General class domains: – Findings: – Events: – Interventions: SUPPQUAL: External data upload: New domain: CO, DM VS AE EX SUPPAE, -BA, -DM, -EX, -LB LB BA Future scope: Trial Design RELREC

CDISC integration system set-up Metadata definition: – Create metadata specifications OC level: – Use CDASH/SDTM variables – Define DCM, DCI and DCI books – CDASH/SDTM annotated CRFs – External data upload RDC level: – Patient data entry OC level: – OC extract creation and output SDTM converter level: – CDISC (CDASH/SDTM) standards library and business rules – SDTM converter mechanism – Generation of SDTM datasets, metadata and reports/warnings

OC level: Use CDASH/SDTM variables Define in Global Library: – DVGs (SDTM controlled terminology) – Questions (CDASH/SDTM variables) – Questions groups

OC level: Define DCM, DCI & DCI books Define at the study level: – DCM – DCI – DCI books 1 DCI -> 1 DCM -> multiple QG Create default view definitions

OC level: set-up rules & naming conventions Kept to a minimum Use CDASH/SDTM variables ONLY except – – Rename &DOMAIN.TESTCD  &DOMAIN.TSTC – Rename &DOMAIN.ORRES  &DOMAIN.ORES All Non CDASH/Non SDTM variables populate SUPP datasets Potential use of extended attributes for numeric values (to produce SDTM+)

Data flow for e.g. Demography

Agenda 1 Introduction 2 CDASH/SDTM in OC/RDC 3 SDTM Converter Mechanism 4 Conclusion

SDTM Conversion – High Level Perspective

Data flow for Demography

Data flow for Adverse Experiences

SDTM converter input: CDISC standards library Used by SDTM converter to interpret standard source or system variables i.e., SDTM or CDASH or OC system – Metadata SDTM variables – Metadata CDASH variables – Metadata OC system variables

SDTM converter input: Short list of business rules Defines generic SDTM conversion choices: USUBJID ОSTUDY||SUBJID ОSTUDY||SITEID||SUBJID Оinclude custom macro RFSTDTC ОMIN (EX.STDTC) ОMIN (DS.DSSTDTC) where MIN (VISITNUM) ОMIN (DS.DSSTDTC) where VISIT = "RANDOMIZATION" Оinclude custom macro RFSTDTC ОMAX (EX.STDTC) ОMAX (DS.DSSTDTC) where MAX (VISITNUM) ОMIN (DS.DSSTDTC) where VISIT = "COMPLETED" or "DISCONTINUED" Оinclude custom macro --DTC О--DAT/--TIM О--DAT ОVISDAT ОVISDAT/VISTIM Оinclude custom macro --BLFL ОBaseline flag = "Y" where MIN --DTC ОBaseline flag = "Y" where MIN (VISITNUM) ОBaseline flag = "Y" where VISIT = "SCREENING" Оinclude custom macro

SDTM converter input: Study parameter paths Define project paths in a Central Parameter Table

SDTM converter input: Short list of default values A default (and editable) value table is used by the SDTM converter DCM NameQGDomainVariableDefault Value EX1EX1EX1EXEXTRT“DRUG A” EX1EX1EX1EXEXDOSE20 EX2EX2EX2EXEXTRT“DRUG B” EX2EX2EX2EXEXDOSE40 ………..

SDTM conversion – high level perspective

ISO 8601 date transformation Rule - Converts variables ending with DAT & TIM in ISO 8601 date format

Automatic transpose for findings Rule - Transposes finding variables TESTCD and corresponding results. Puts records of variables ending with the same number together

Global transformations Rule - Compute values based on rules in business rule sheet USUBJID ОSTUDY||SUBJID ОSTUDY||SITEID||SUBJID Оinclude custom macro

Domain specific transformations Rule - Populate variables with values specified in default value sheet EXDOSE EXTRT EXDOSE EXTRT Variable 40EXEX2EX2EX2 “DRUG B”EXEX2EX2EX2 20EXEX1EX1EX1 “DRUG A”EXEX1EX1EX1 Default Value DomainQGDCM Name

Generate SEQ variable Rule – Generate SEQ variable based on Keys specified in Table Level Metadata in the CDISC Standards Library Rule - For custom domains use the following: – Event: --TERM, -- STDTC – Intervention: --TRT, --STDTC – Finding: --TESTCD, VISITNUM, TPTNUM, --DTC

Extract TEST from DVG for TESTCD Rule - Populates TEST by extracting DVG long value of TESTCD Vertical structure: Horizontal structure:

Create Supplementary Datasets Rule - Non-CDASH/Non-SDTM/Non-OC System variables to populate supplementary datasets

Create Comment Dataset Rule - Collect comments in OC in &DOMAIN._COVAL variable

Create SDTM Dataset Rule – Use variable level attributes from the CDISC standards library

SDTM conversion – high level perspective

SDTM converter output: Reports and Warnings Data profile reports of the SDTM generated datasets.

SDTM converter output: Reports and Warnings Warning reports: includes basic checks for conformance with the CDISC domain models: – Including all required and expected variables as columns in a domain – Using CDISC-specified standard variable names – Using CDISC-specified data types for all variables – Adhere to SAS v5 XPT limitations These conformance checks are a subset of the complete set of CDISC SDTM compliance checks

SDTM converter output: Metadata Table Metadata Variable Metadata Controlled Terminology Value Level Metadata Dataset Name Dataset Label Class Structure Purpose Keys Variable Name Variable Label Type Length Controlled Terminology Role Source Variable Value Label Type Controlled Terminology Code Value Code Text

Agenda 1 Introduction 2 CDASH/SDTM in OC/RDC 3 SDTM Converter Mechanism 4 Conclusion

Maximize use of CDASH and SDTM variables Use CDISC controlled terminology Through smart set-up in OC, automate SDTM generation through a generic SDTM converter mechanism – No SDTM conversion required in OC/RDC; occurs outside OC/RDC – No additional programming is required – Real-time SDTM datasets and metadata, ready for analysis

Peter Van Reusel Business Unit Director CRO Services Business & Decision Life Sciences Tel Fax Mobile Sint-Lambertusstraat 141 Rue Saint- Lambert 1200 Brussels Thank you for your attention